Feb 28, 2010

Cloud Storage API(s)

I am not sure whether we should call this as innovation or insanity!

There are 13 different cloud storage API's in the market today. Here is the list and you can click on each one of them for the details.
  1. Amazon S3
  2. Nirvanix
  3. SNIA-CDMI
  4. AT&T
  5. Simple Cloud
  6. Cloud Loop
  7. Iron Mountain
  8. Rack Space
  9. Box.Net
  10. JClouds
  11. Google DataStore
  12. Microsoft Azure
  13. EMC Atmos
  14. Para Scale & Zetta: Not sure whether they also have proprietary API
I don't know if I missed any thing!

I am also not including any key-value pair storage products here. There is an effort behind each one of them and simply can't be ignored. The cloud storage(exposing storage over HTTP) is real. We need a standard.

Feb 19, 2010

App Cloud Triangle

We all know that how important is the developer community in the computer industry. An example: Click here to see the famous video of Steve Ballmer, shouting developers,developers,...

In the last 2 decades, companies like IBM, Borland, Apple, Microsoft, Sun and later open source foundations like Apache was driving the dev community.(mainframes, client/server & web)

Now the landscape is changing again! This time it's cloud. If you have any question about the success of cloud,just take a look at RightScale website. It says as of today, launched more than 900,000 cloud servers within short period of time! Today complete web platforms are built exclusively on cloud w/o internal data-centers!

So, the question is who is going to drive the web developer community in the cloud? The answer is whoever going to be the force behind the application ecosystem in the cloud. It may change in the future but now, I see that the following are the 3 main players going to be driving the developer community in the cloud.

Feb 5, 2010

What is Scala?

This is yet another blog to demystify a technology. Scala is yet another experiment may/may not survive in the long run. I do see that there is some level of interest in this language. I was trying to find out what is it & what is not. More specifically, why it got the attention it deserves.

Ruby is Out. In comes Scala:
Twitter is using Scala now because it fell flat on the face with Ruby. (do a Google search if you want more explanation).

Ok, fine.

What is Scala?
It's Java!

What?
Yes, it gives you a different syntax but it compiles into Java byte-code (.class file) and runs on top of Java Virtual Machine.

Then, why can't I use Java itself?
I don't know. My head hurts if i start reading the list about Scala over Java. How can it be? Scala just produces Java class file. Sure, there may be some convenience factor in the syntax. There is no technical or business merit.

So, what is Scala? (technical-definition)
Scala = Object Oriented Programming + Functional Programming.
I don''t like it. This is like combining Ice Cream with Hot Coffee. If a technology/product combines 2 different discrete things then it has no merit. It's merely trying to justify it's existence.

Add to the confusion, i read about it's compatibility with C#. Good luck!

What is Scala one more time?
It's like driving to San Francisco from Sunnyvale. You can take Highway 101 or Highway 280. You have a choice but end result is same. It's SF!

Pl don't get me wrong- i may end up using Scala, sorry Java!

Feb 2, 2010

What is Hadoop?

What is Hadoop? (plain-English)
Hadoop is a Java framework which brings compute to the storage.

What is Hadoop? (technical-definition)
Hadoop = HDFS + MapReduce.

What is HDFS?
HDFS=Hadoop Distributed File System.

You can think of this way-
it takes a single file, cuts into multiple chunks and distributes it across many computers. Its an layer on top of disk file system.

What is MapReduce?
Distributed processing framework, works on top of HDFS.

You can this of this way-
after HDFS distributes data to different computers, MapReduce sends programs to those computers to work on those data.

==================================
Here is the reason for this blog:

I recently attended "an introduction to Hadoop" talk and this event attracted closer to 200 people or so. Including the organizer of the event, everyone was surprised to see these many people coming to this event. I thought either Hadoop user-base is expanding OR still people are trying to understand what is Hadoop & how can i use it? It looks to me that still Hadoop is in early stage & the community is trying to learn it.

The presenter did a great job explaining the basics of Hadoop. Thank you. I cut the verbose out from the presentation and posting the core of the presentation below. I hope now you should be able to understand the basics of Hadoop.

===================================